An Improved Decomposition Multiobjective Optimization Algorithm with Weight Vector Adaptation Strategy

Multiobjective evolutionary algorithms based on decomposition (MOEA/D) decomposes a multiobjective optimization problem (MOP) into a number of scalar optimization subproblems, and solves each subproblem in a collaborative manner. For a MOP, different subproblems often have different difficulty to be approximated, especially when the MOP is extremely complex or has a discontinuous optimal front. This paper proposes a weight vector adaptation strategy for this issue, which changes the weight vectors and optimizes their computational resource allocation to fit the MOP. The experimental results on a variety of MOP test instances show that the proposed algorithm is competitive in comparison with three state-of-the-art decomposition multiobjective evolutionary algorithms.